mechanical analysis of controls on strain partitioning in the...

Mechanical analysis of controls on strain partitioningin the Himalayas of central Nepal

V. Godard1,2 and D. W. Burbank2

Received 29 January 2011; revised 20 June 2011; accepted 7 July 2011; published 7 October 2011.

[1] We present a mechanical analysis of the problem of slip partitioning between themajor thrust systems in a collisional range. We focus on two structures in the Himalayasof central Nepal: the Main Himalayan Thrust (MHT) and the Main Central Thrust (MCT).We use finite element modeling to test the influence of various parameters, such asfriction coefficients and surface processes, and we investigate how they affect thedistribution of deformation between these two faults. We observe that reproduction of thelate Quaternary kinematic pattern across the range with our model requires strictconditions on the friction coefficients, such that the MHT is very weak, whereas the MCTis significantly stronger. The most important parameter that controls slip partitioningappears to be the dip angle of the MCT, with a gentler or steeper MCT promoting orinhibiting slip, respectively. We also show that transient loading and unloading throughfocused glacial erosion in the higher part of the range can unclamp the MCT and allow asignificant increase in slip rates. The results of this mechanical sensitivity investigationhave important implications for the dynamics of the Himalayan wedge and point towardalong‐strike structural variations as a first‐order control on slip partitioning.

Citation: Godard, V., and D. W. Burbank (2011), Mechanical analysis of controls on strain partitioning in the Himalayas ofcentral Nepal, J. Geophys. Res., 116, B10402, doi:10.1029/2011JB008272.

1. Introduction

[2] Active mountain ranges comprise rapidly evolvingstructural complexes whose behavior is either controlled byinternal factors, e.g., lithospheric rheology, or external ones,e.g., intensity of erosion and sedimentation. One of adeforming range’s most salient characteristics and probablythe most important when it comes to understanding itskinematics is the existence of crustal‐scale faults thataccommodate a large fraction of the far‐field strain. Doc-umenting the timing and magnitude of slip that is accom-modated on those individual structures has the potential toyield critical insights on the dynamics of the whole orogen[e.g., Lavé and Avouac, 2001; DeCelles et al., 2001].Extensive studies have been carried out in the majormountain ranges around the globe and, in many settings,have yielded a relatively thorough documentationconcerning (1) the architecture of the range [Martin et al.,2005], (2) the boundary conditions to which it is submitted,e.g., the magnitude of erosion and far‐field long‐termvelocities [Bettinelli et al., 2008], and (3) its internaldeformation pattern [Le Fort, 1975; Schelling and Arita,1991; Harrison et al., 1998; Hodges, 2000; DeCelles et al.,2001; Avouac, 2003; Robert et al., 2009]. A process‐

based understanding of the dynamics of those orogenicsystems that links all the observations mentioned above isan attractive research objective and lies at the core ofnumerous studies. However, a mountain range’s actualphysical behavior, which directly controls the way appliedboundary conditions are converted into an observeddeformation pattern, is still poorly understood. In particu-lar, the mechanical principles for the partitioning ofdeformation across multiple fault systems are still largelyunresolved, largely because of the difficulties of investi-gating changing fault behavior over different timescales[Friedrich et al., 2003].[3] With respect to orogenic processes, a key result of the

last two decades of research is the conceptual developmentof the idea of interactions and couplings between tectonicsand surface processes [Beaumont et al., 1992; Avouac andBurov, 1996; Willett, 1999; Thiede et al., 2004; Whippleand Meade, 2006; Tomkin and Roe, 2007; Berger et al.,2008; Whipple, 2009]. Although the primary factors con-trolling strain patterns in mountain ranges were widelyrecognized to be the rheology of the lithosphere, faultstrength or the characteristics of the far‐field kinematicforcing, several studies have concluded that surface pro-cesses could also focus deformation in regions of higherosion and, as a consequence, control the strain pattern ofthe whole orogen [Willett, 1999]. Those surface processesactually encompass a wide spectrum of phenomena rangingfrom various modes of large‐scale erosion [Beaumont et al.,2001] to the filling of peri‐orogenic sedimentary basins[Fuller et al., 2006] or the growth and retreat of ice bodies

1CEREGE, CNRS, UMR 6635, Aix‐Marseille Université, Aix‐en‐Provence, France.

2Earth Research Institute, University of California, Santa Barbara,California, USA.

Copyright 2011 by the American Geophysical Union.0148‐0227/11/2011JB008272

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 116, B10402, doi:10.1029/2011JB008272, 2011

B10402 1 of 19

http://dx.doi.org/10.1029/2011JB008272

[Hampel et al., 2007]. These processes nevertheless share acommon mechanical implication which is to add or removemass within the range and, as a consequence, to significantlyperturb the crustal stress field. A particularly importantprevious result has been the numerical prediction that suchsurface processes can not only influence the large‐scale andlong‐term deformation of mountain ranges, but that theyalso have the potential to modulate the slip rate of individualfaults at timescales of a few thousands years [Hetzel andHampel, 2005; Hampel and Hetzel, 2006; Hampel et al.,2007; Turpeinen et al., 2008; Hampel et al., 2009;Maniatis et al., 2009]. Such results hold first‐order im-plications for the understanding of the partitioning in spaceand time of shortening between the major faults of an activeorogenic wedge.[4] The purpose of the present study is to use a mechan-

ically based approach to investigate the factors controllingthe partitioning and accommodation of deformation in anactive orogenic system. Observations from the Himalaya ofcentral Nepal underpin this exploration. The Himalayashave been a propitious site for the study of continentaltectonics and geomorphology, including the investigation ofpotential coupling between the two, because of the intensityof climatic and kinematic gradients across the range. Thisrange is also of particular interest for our study becausekinematic reconstructions suggest that several crustal‐scalethrust faults may have been sequentially active over the lastfew million years [Schelling and Arita, 1991; Schelling,1992; DeCelles et al., 2001; Robinson et al., 2003; Searleand Godin, 2003; Pearson and DeCelles, 2005; Leloup et al.,2010]. However, the mechanical rationale for the partition-ing of deformation between these faults is only partiallyunderstood and is, therefore, the main question addressed bythis study.[5] First, we present the geodynamical context of the

Himalayas of Central Nepal with a focus on aspects that arerelevant to the specific problem of the shortening distribu-tion between the major thrust faults. Next, we introduce theprinciples of the geodynamic modeling approach that we useto tackle this problem. Then, after describing the resultsassociated with our reference model, we successivelyexplore and discuss a range of internal parameters andexternal forcings that are likely to have an influence on thepattern and magnitude of deformation partitioning.

2. Tectonic Setting and Problem Statement

2.1. Overview of the Himalayas of Central Nepal

[6] We focus here on data that are specifically relevant toour study and refer the reader to published reviews for amore complete coverage of the geological and geodynami-cal setting of the Himalayas of Central Nepal [Gansser,1964; Le Fort, 1975; Hodges, 2000; Avouac, 2003; Yin,2006; Avouac, 2007]. As one of the main expressions ofthe ongoing collision between India and Eurasia, theHimalayan range currently accommodates ∼30–50 % of theconvergence between the two plates [Bilham et al., 1997;Jouanne et al., 1999; Larson et al., 1999; Zhang et al.,2004; Bettinelli et al., 2006]. From a morphological pointof view, the Himalayas also constitute one of the steepesttopographic escarpments on Earth with a transition fromnear sea level in the Gangetic plain to ∼5000 m on the

Tibetan Plateau occurring across an horizontal distance of∼120 km.[7] From south to north, the range comprises a distinct

succession ofmorpho‐tectonic units that are bounded bymajorcrustal‐scale, mostly contractional, structures (Figure 1). Themost external unit is the Siwaliks range that is activelydeforming in the hanging wall of the Main Frontal Thrust(MFT) and is overthrusting the Gangetic foreland. The MFTis currently the most active structure of the Himalayan rangeand, at least in parts of central Nepal, it accommodates mostof the 20 mm/yr of Holocene shortening between India andsouthern Tibet [Lavé and Avouac, 2000]. The Siwaliksfoothills are bounded on their northern edge by the largelyinactive Main Boundary Thrust (MBT) that marks thetransition to the Lesser Himalaya. Spanning from the MBTto the Main Central Thrust (MCT), the Lesser Himalaya(LH) is typified by relatively subdued topography and lowrock uplift rates, and it mainly comprises metasediments ofIndian affinity [Bollinger et al., 2004a]. Topography starts torise significantly when approaching a physiographic transi-tion (PT2: [Catlos et al., 2001; Hodges et al., 2001; Wobuset al., 2003]) and entering the metamorphic and crystallinesequence of the Higher Himalaya (HH) that lies north of theMCT and is characterized by steep hillslopes and entrenchedgorges [Pratt et al., 2002; Gabet et al., 2004a]. Finally, theSouth Tibetan Detachment system separates the southernHimalayan units from the Tethyan Sedimentary Series thatconstitutes a large part of the high range [Searle and Godin,2003]. Underlying the Lesser and Greater Himalaya andconnecting with the MFT, the Main Himalayan Thrust(MHT) is considered to be the plate boundary interfacealong which India underthrusts the Himalayas [Zhao et al.,1993; Nabelek et al., 2009].[8] Even if the MFT is the main active structure at present,

the late Cenozoic history of the Himalayas is characterizedby a complex sequence of southward activation and aban-donment of the major structures [Leloup et al., 2010].Although the central Nepal region of the Himalayas is one ofthe most intensively documented mountain ranges, numer-ous controversies and unresolved questions concerning itsgeodynamical and geomorphological activity remain unre-solved and lie at the core of active investigations.

2.2. Glaciations in the Himalayas of Central Nepal

[9] Glaciers are widespread in high elevation areas in theHimalayas of central Nepal and can occupy a significantfraction of the landscape above 5000 m [Pratt‐Sitaula,2005; Harper and Humphrey, 2003]. Glaciers display awide range of variations in their present ice‐discharge pat-tern and surface evolution along the Himalaya arc. Suchvariations can be related to a dual influence on theirdynamics and mass balance that depends on both globalnorthern hemisphere temperature and the intensity of theAsian monsoon. However, local geomorphic factors and, inparticular, contrasts in supra‐glacial debris cover also have acritical influence [Scherler et al., 2011]. Few studies haveattempted to document the current effective erosion that canbe attributed to glaciers in the Himalayas. A local study inthe Marsyandi in central Nepal catchment shows that, onaverage, glacial erosion rates are <10 mm/yr and are in partcontrolled by orographic precipitation with rates <5 mm/yrin the dry northern part of the catchment [Godard et al.,

GODARD AND BURBANK: STRAIN PARTITIONING IN THE HIMALAYAS B10402B10402

2 of 19

Figure 1. (a) Generalized map of the Himalayas of central Nepal, showing the position of the mainfaults discussed in this study. The Marsyandi catchment (black box, inset b) and the cross section usedto project the data (insets c‐h) are also indicated. (b) Overview map of the Marsyandi catchment in centralNepal, showing the main tectonic units [Searle and Godin, 2003] and modern glaciers [Armstrong et al.,2005]. Black lines denote the extent of early Holocene glaciers proposed by Pratt‐Sitaula [2005]. Blackbox indicates the location of Figure 2. Cross sections (N18°) across central Nepal displaying different datasets available. (c) Schematic geological cross section [Lavé and Avouac, 2001; Avouac, 2003; Hodgeset al., 2004] corresponding to the geometry used in the mechanical model (Figure 3). MCT: MainCentral Thrusts; MFT: Main Frontal Thrust; MHT: Main Himalayan Thrust. (d) GPS velocities (circles)in a fixed India reference frame [Bettinelli et al., 2006] and longer term estimates for the convergenceacross the Himalayas (squares: Holocene [Lavé and Avouac, 2000], Quaternary [Armijo et al., 1986] andLate Cenozoic [Lyon‐Caen and Molnar, 1985]). (e) Topography (mean, ±1s and extremal values) andprecipitation pattern derived from TRMM data [Bookhagen and Burbank, 2006] and station network[Burbank et al., 2003]. (f) ±1s topographic envelope (dark gray) and areal percentage of modern glacialcoverage (red line) [Armstrong et al., 2005]. (g) Synthetic fluvial incision profile derived from shear‐stress analysis across central Nepal [Lavé and Avouac, 2001]. Erosion rates inferred from mineralogicalspectra analysis [Garzanti et al., 2007] and sediment gaging [Gabet et al., 2008]. The light blue and reddomains indicate the extents of two distinct geomorphic domains arising from the comparison of short‐[Gabet et al., 2008] and long‐term [Blythe et al., 2007] erosion estimates. See text for details. (h) Fissiontrack ages in the Marsyandi catchment [Burbank et al., 2003; Huntington et al., 2006; Blythe et al., 2007].Note that the horizontal scale is different for this inset. Light blue and red domains are the same as inprevious inset.


3 of 19

2011]. Important fluctuations in the extension of Himalayanglaciers have been documented over the Quaternary [Owenet al., 2002; Owen and Benn, 2005; Gayer et al., 2006;Owen et al., 2008]. In central Nepal, the depression of theLast Glacial Maximum (LGM) equilibrium line altitude(ELA) has been estimated to be between 500 and 1000 m[Duncan et al., 1998; Harper and Humphrey, 2003; Pratt‐Sitaula, 2005], which suggests likely changes in the rela-tive importance of glaciers in the erosion of the Himalayanrange. However, no estimates of the magnitude of pastglacial erosion rates currently exist in the area, althoughfar‐field estimates suggest that the global sediment flux outof the range was significantly modulated by variations inglacial coverage [Goodbred and Kuehl, 1999; Rahamanet al., 2009].

2.3. Two Himalayan Open Questions

[10] In Central Nepal, recent reconstructions of tectonicand erosive processes in space and time highlight twoimportant unresolved questions: to what extent is Himalayanshortening partitioned onto faults other than the MainFrontal Thrust; and can climatically modulated variations inerosion and in both sediment and ice loading affect slip rateson major faults?2.3.1. Quaternary Kinematic Evolution of CentralNepal[11] Current geodetic deformation rates [Bilham et al., 1997;

Larson et al., 1999; Jouanne et al., 1999; Zhang et al., 2004;Bettinelli et al., 2006] and longer‐term studies [Lyon‐Caen andMolnar, 1985; Armijo et al., 1986; DeCelles et al., 2001]indicate that shortening across the central Himalayas in Nepalis ∼2 cm/yr. For the Holocene period, a study of deformedterraces at the front of the range lying south of Kathmandushowed that most, if not all, of this shortening between Indiaand southern Tibet could be accounted for by slip on the MFTin the Himalayan foreland [Lavé and Avouac, 2000]. At thescale of the whole range, this observation has spurred thedevelopment of a widely recognized geodynamical and seis-motectonic model for the Himalayas of central Nepal in whichthe MHT is the major active crustal fault and its ramp‐flatstructure is responsible for the observed interseismic and long‐term patterns of rock uplift [Lavé and Avouac, 2001; Cattinand Avouac, 2000; Avouac, 2003; Bollinger et al., 2004b].

[12] In contrast, west of the Kathmandu transect, recentstudies have documented offsets of both apatite fission trackand muscovite 39Ar‐40Ar ages (Figure 2 [Huntington andHodges, 2006; Blythe et al., 2007]), as well as structuraland morphological studies [Hodges et al., 2004] at somelocations across the MCT Zone, that suggest that a modestbut significant fraction of the Indo‐Tibetan convergencecould have been absorbed by slip on the MCT since 2.5 Ma.Those observations have underpinned a geodynamicalmodel of central Nepal in which out‐of‐sequence thrustingon the MCT significantly contributes to the kinematicbudget of the range and has moved the question of a LateCenozoic reactivation of the MCT to the center of ongoingdebates [Bollinger et al., 2004a; Hodges et al., 2004;Bollinger et al., 2004b; Wobus et al., 2005; Whipp et al.,2007; Robert et al., 2009; Herman et al., 2010a; Robertet al., 2011].[13] One possible way to reconcile these divergent

observations (all slip occurs on the MFT versus significantslip on the MCT) would be to have intermittent activity onthe MCT, with the last active episode being pre‐Holocene[Hodges et al., 2004]. Alternatively, the discrepanciesbetween those observations and interpretations could resultfrom significant along‐strike variations in the structures ofthe Higher and Lesser Himalayas in central Nepal, espe-cially as related to the complex structure of the MCT in thatregion and the existence of the massive, but laterally limitedKathmandu klippe (Figure 1a). In any case, a betterunderstanding of the mechanical behavior of the Himalayanwedge has the potential to deliver critical insights on thedynamics of those different faults and the parameters thatcontrol slip partitioning across the range, which is the mainpurpose of the present study.2.3.2. Time Variations of Erosion in the High Range[14] Another pending question and interesting conundrum

in the Himalayas relates to the variation of erosion in spaceand time over climatic timescales. Almost 20 years ofextensive studies in Central Nepal, particularly in the Mar-syandi catchment, have documented spatial patterns in sur-face processes in significant detail and across severaltimescales [Lavé and Avouac, 2001; Burbank et al., 2003;Gabet et al., 2004b, 2004a; Huntington and Hodges, 2006;Huntington et al., 2006; Brewer et al., 2006; Blythe et al.,2007; Garzanti et al., 2007; Gabet et al., 2008, 2010].

Figure 2. Sample location map and age‐elevation profiles for the low‐temperature thermochronologydata available across the MCT zone in the Marsyandi catchment, displaying an offset in colling ages thatcould be a manifestation of Quaternary reactivation on related structures [Huntington et al., 2006; Blytheet al., 2007]. See Figure 1b for location.


4 of 19

Gradients in modern erosion rates and precipitation seem toindicate that erosion in the Himalayan rain shadow[Bookhagen and Burbank, 2006] is up to fivefold slowerwhen compared to wetter areas to the south [Garzanti et al.,2007; Gabet et al., 2008] (Figure 1g, light blue (dry) and red(wet) domains, respectively). In contrast, an exceptionallydense data set of apatite fission track ages in the same regionsuggests that at ≥105‐yr timescales, erosion rates are aboutas fast in the currently dry region as they are where themonsoon is dominant [Burbank et al., 2003; Blythe et al.,2007] (Figure 1h). Given that the northern, semi‐arid areahosts the majority of the glaciers today and was much moreextensively glaciated in the past, it has been hypothesizedthat erosion in the currently more arid regions was severalfold faster when glaciers were more expanded in the pastand compensated for slower rates like those of today [Gabetet al., 2008]. In addition, these changes are likely to becoupled to the evolution of the strength of the Indianmonsoon.[15] Recent studies that tie slip on thrust faults to glacial

unloading [Turpeinen et al., 2008] prompt us to examinewhether rapid changes in erosion and ice loads in an arealocated in the hanging wall of the MCT could have had aninfluence on the activity of this fault, as well as whether thedistribution of slip between the major Himalayan thrusts canbe partially modulated at climatic timescales by the locusand rate of glacial erosion and ice loading/unloading.[16] The two potentially related geological problems that

we present here (slip partitioning and the role of climaticallymodulated loads) are research themes that may serve toimprove our understanding of mountain building in generaland Himalayan geodynamics in particular. Using those ob-servations and open questions as a starting point, this studyattempts to address some basic, but still unresolved pro-blems about the mechanical behavior and sensitivity of theHimalayan wedge and about the controls on slip partitioningbetween major faults in active tectonics settings.

3. Modeling Approach

[17] Here we present the rationale of our modelingapproach and the main constitutive components of themechanical model we use. Most of the pertinent ideas havebeen previously presented in papers to which we refer forfurther discussion of the numerical approach [Cattin andAvouac, 2000; Godard et al., 2004, 2006, 2009]. The aimof our modeling approach is not to attempt to fit a specific

pattern or sequence of events for the Himalayas of centralNepal, but rather to try to gain a general understanding ofthe mechanical behavior and sensitivity of the orogen todifferent parameters and forcings. We investigate processesat a timescale of 100 ky, and our modeling is based on theavailable knowledge for the present structure and deforma-tion of the range. We look at the dynamics of the rangestarting from specified initial conditions, and we do notinvestigate here the long‐term evolution of the Himalayas orthe emplacement of the main structures [Beaumont et al.,2001], but rather focus on an instantaneous snapshot ofthe mechanical behavior, given a configuration that attemptsto mimic the available geophysical and geological data.

3.1. Mechanical Principles and Numerical Approach

[18] The mechanical problem is solved using the finiteelement method with a time integration based on thedynamic relaxation algorithm [Underwood, 1983; Hassaniet al., 1997]. The rheology of the lithospheric materials isconsidered to be visco‐elasto‐plastic (Table 1). Plasticityis modeled using a Drucker‐Prager formulation. Viscosity isconsidered to be controlled by a non‐linear and thermallyactivated flow law:

_" ¼ �0 �1 � �3ð Þne�EaRT ; ð1Þ

where _" is the strain rate, T the temperature, R the universalgas constant, and s1 and s3 are the maximum and minimumprincipal stresses, respectively. The controlling parametersare listed in Table 1. Classical Coulomb friction is used tosimulate the evolution of the contact interfaces. We refer toearlier studies for an in‐depth description of the physicallaws implemented and the numerical approach used to solvethe constitutive equations along with their associated sta-bility issues [Hassani et al., 1997; Godard et al., 2009]. Thesimulation duration is 500 ky of which the first half isdevoted to an initial phase of mechanical stabilization of thesystem. Tectonic and erosion processes are graduallyintroduced after this initial stabilization period. The typicaltime step is 0.1 year. This allows us to investigate processesat a time‐scale and a resolution that are relevant for both theHolocene deformation of the range and longer term evolu-tion over the Quaternary. Due to the relatively short timespan of our simulation, we do not solve the heat equationover time: instead we use a fixed thermal field [Henry et al.,1997]. One important aspect of this thermal field is that itintroduces a high‐temperature and low‐viscosity zone under

Table 1. Reference Rheological Parameters Used for the Visco‐elasto‐plastic Formulation Used in This Studya

Upper Crust (dry quartzite) Lower Crust (Maryland diabase) Mantle (dry olivine)

ElasticityYoung modulus, E (GPa) 50 50 150Poisson’s ratio, n 0.25 0.25 0.25Density, r (kg/m3) 2900 2900 3300

PlasticityCohesion, c (MPa) 10 10 10Internal friction angle, � 30° 30° 30°

ViscosityStandard fluidity, g0 (Pa−n/s) 6.3 10−24 6.31 10−20 7.9 10−18

Power law exponent, n 2.72 3.05 3.5Activation energy, Ea (kJ/mol) 184 276 528

aKirby [1983], Carter and Tsenn [1987], Kirby and Kronenberg [1987] and Tsenn and Carter [1987].


5 of 19

the high range and southern Tibet that acts as a crustal‐scaledecoupling level and allows the transmission of the kine-matic boundary conditions to the main faults (MCT andMHT).

3.2. Geometry of the Model and Boundary Conditions

[19] The general geometry of the model is similar to thatused in previous works [Cattin and Avouac, 2000; Godardet al., 2004, 2006] (Figure 3) and distinguishes between therheological properties of the upper and lower crust [Godardet al., 2004]. The geometry mimics the main features of theHimalayan orogen in central Nepal [Zhao et al., 1993;Cattin et al., 2001; Hetényi et al., 2006; Nabelek et al.,2009] and is specifically designed to match the cross sec-tion proposed by Hodges et al. [2004]. We use triangularelements with an average characteristic dimension of 4 km.An explicit description of the MCT as a frictional interfaceis a new feature in this class of Himalayan models.[20] Vertical displacements are free, whereas horizontal

displacements are set to zero on the vertical faces of themodel, except for the area where a 20‐mm/yr convergencevelocity is applied (Figure 3). A hydrostatic foundation isintroduced at the base of the model to account for isostaticsupport.

3.3. Erosion Formulation

[21] In order to ensure that an approximate topographicsteady state condition is attained and to avoid multiplying thenumber of free parameters, we do not use a functional rela-tionship to simulate background erosion [Avouac and Burov,1996; Willett, 1999; Godard et al., 2004], but rather enforcea surface condition whereby all tectonically induced rock‐uplift is compensated by erosion. This constraint provides apurely kinematic definition of erosion that does not allow forfull feedback mechanisms between tectonic and erosion todevelop. This specification defines the background erosionacting on the topography upon which will be superimposed,for some simulations, shorter wavelength time‐varying ero-sion to test the sensitivity of the range to unloading in spe-cific areas. Coupled thermo‐kinematic and surface evolution

modeling supports the idea that the long‐wavelengthtopography is close to steady state in central Nepal [Hermanet al., 2010a].

4. Slip Partitioning and Friction Coefficients

[22] Our reference model is used as a baseline againstwhich to analyze the sensitivity to different types of per-turbations. A range of friction coefficients is explored forboth the MCT and MHT. Slip rates were calculated for bothfaults at the end of the simulation (Figures 4a and 4b).[23] For the range of friction coefficients explored here,

the system displays high variations in slip‐rate values, cor-responding to highly contrasting behaviors that range frommost of the convergence being accommodated by the MHTfor high values of the friction coefficient on the MCT andlow values on the MHT, to the opposite configuration. Thistrade‐off is a simple expression of the minimum workprinciple that favors the fault which is the most energeticallyefficient in accommodating deformation. If we consider theHolocene kinematic situation in central Nepal, in whichmost of the convergence appears to be accommodated bythe MHT and limited slip occurs on the MCT, such slippartitioning points to the existence, in our models, of arelatively narrow range of friction coefficients for both theMHT and MCT that allows us to reproduce such observa-tions (shaded region on Figures 4a and 4b). The current slippartitioning, therefore, implies a strong MCT (m > 0.1) thatminimizes the amount of slip accommodated on it, whereasthe MHT is weak (m < 0.02) such that most of the conver-gence occurs on this thrust. Such low values for the frictioncoefficient on the MHT are consistent with some estimates ofthe friction on major continental faults [He and Chéry, 2008].We also note that, within the shaded zone (Figures 4a and4b), the slip rates for both the MCT and MHT are highlysensitive to changes in the friction coefficient. Such gradientssuggest that small variations of the friction coefficients oneither the MCT and MHT could lead to significant mod-ifications of the slip partitioning between those two faults.

Figure 3. Structure of the mechanical model and boundary conditions used in this study. Note that thethermal field is considered static [Henry et al., 1997] and that the heat equation is not solved. The thick,dark gray contours indicate the 500° and 700°C isotherms; thinner dashed contours are 100°C incrementisotherms. The structure is supported by a Winkler foundation simulating isostasy. Vertical faces arepermitted free vertical displacements, but no horizontal displacements. Dashed lines on both sides of theMCT indicate the laterally shifted positions of the MCT used in Figure 6. Horizontal black line denotesthe portion of the topography that is subject to focused loading or unloading in Figures 8 and 9. The blackdashed box indicates the position of Figure 1c.


6 of 19

This model response implies a potentially high sensitivity ofthe system to perturbations.[24] The difference between the applied far‐field conver-

gence rate and the slip rates actually accommodated on theMCT and MHT provides an estimate of the amount ofconvergence that is not taken up on the MHT/MCT system,but is instead expressed as distributed strain and crustalthickening (Figure 4c). Such off‐fault deformation is highestwhen the friction coefficients are large and thus when themechanical resistance of the fault system as a whole ismaximized. Slightly negative values for low MHT friction

(Figure 4c) are indicative of a small amount of gravitationalcollapse of the topographic margin toward the foreland andare mainly promoted by a weak MHT. For such low frictioncoefficients on the MHT, the sensitivity to changes in thefriction on the MCT is decreased. Such behavior is consis-tent with a predominance of large‐scale displacement andgravitational collapse by basal sliding on the MHT, whereasthe MCT has only a second‐order mechanical role.[25] For selected friction coefficients, individual time

series of the finite slip on the MCT and MHT show theimportant role of the MCT friction in modulating the slip of

Figure 4. Exploration of the influence of different combinations of friction coefficients for the MCT andMHT in terms of slip rates on those structures. (a) Slip rate on the MCT. Light gray areas indicate therange of values for the friction coefficients that can be considered as reasonable given the Quaternarykinematic pattern in central Nepal (slip on the MCT ranges between 0.5 and 4 mm/yr and slip on theMHT is greater than 16 mm/yr). The contoured surface is derived by griding from more than 2000 indi-vidual model runs (small dots), sampled in the parameter space [Wessel and Smith, 1991]. (b) Slip rate onthe MHT. (c) Difference between the applied velocity (20 mm/yr) and the combined slip rates on theMCT and MHT. This difference is considered as an estimate of the amount of shortening that isaccommodated by distributed crustal deformation. Overall, slip rates increase on the MCT for lowerfriction on it and/or for higher friction on the MHT. Increasing friction on both faults leads to increasedoff‐fault deformation. (d–g) Finite slip evolution on the MHT and MCT for selected combinations offriction coefficients.


7 of 19

the whole fault system (Figures 4d–4g). Variations of theMCT friction from 0.1 to 0.2 imply a transition from apartitioning pattern where both faults have roughly equalcontributions in the accommodation of deformation towarda situation of significantly suppressed slip on the MCT anddominant activity on the MHT.

5. Internal Controls on Shortening Partitioning

[26] We investigate here how some internal parameters,such as structural configuration, influence the partitioning ofdeformation between the MCT and MHT. Fault geometrydirectly influences how the ambient stress field is expressedas shear and normal stresses on the fault plane and, there-fore, plays an important role in promoting or inhibiting slip.At least in some parts of the Himalayas, the MCT is acomplex set of thrust faults instead of a single structure[Hodges et al., 2004; Martin et al., 2005; Searle et al.,

2008]. This network of faults and large‐scale shear zonesdefines the MCT zone which makes the practical definitionof the actual surface trace of the fault difficult. Similarly,whereas the foliation inside the MCT zone has beenextensively measured, the actual dip of the fault plane itselfis still poorly resolved due to the lack of sufficiently high‐definition geophysical imagery. Hodges et al. [2004]reported dip angle values ranging from 19° to 33° for thedifferent thrusts that they mapped in the physiographictransition (PT2) area along the Marsyandi. Other studies incentral Nepal reports values of the foliation in the areaaffected by MCT deformation from 17° to 45°, with mostobservations in the 20–30° range [DeCelles et al., 2001;Hodges et al., 2004; Martin et al., 2005; Pearson andDeCelles, 2005; Martin et al., 2010]. Here we explore theresponse of the system to different changes in the configu-ration of the MCT for both its dip and position with respectto the High Himalayan topographic front. Given the absence

Figure 5. Influence of the dip of the MCT on the slip partitioning between the MCT and the MHT.Plotted values (mm/yr) represent the difference between a model with a steep MCT (30°) and the refer-ence model (20° in Figure 4). (a) Difference in slip rate on the MCT with respect to the reference model(Figure 4a). (b) Difference in slip rate on the MHT with respect to the reference model (Figure 4b).(c) Difference between the applied velocity (20 mm/yr) and the combined slip rates on the MCT andMHT, with respect to the reference model (Figure 4c). (d) Difference in rock uplift rate above theMCT (in %) with respect to the reference model. Slip rates and rock uplift are highly sensitive tofault dip such that higher dips depress slip rates and promote slip transfer to gentler faults.


8 of 19

of data to indicate otherwise, we model the MCT as a planar,uniformly dipping fault surface.

5.1. Dip Angle of the MCT

[27] A striking difference in slip rate occurs between amodel where the dip angle of the MCT is 30° versus thereference model where the dip angle is 20° (Figure 5). Thegeneral outcome of this sensitivity test is that a steeper anglefor the MCT makes it mechanically harder to accommodateconvergence on this structure and, therefore, promotes slipon the MHT. As could be intuitively expected, in a steeperconfiguration, a larger amount of the horizontal stress re-sulting from the boundary conditions is going to be ex-pressed as normal stress on the fault plane and that stressinhibits slip. The difference with respect to the referencemodel is least pronounced for a combination of high frictionon the MCT and low friction on the MHT. Indeed, suchfriction coefficients tend to drastically diminish themechanical role of the MCT and make the system globallyinsensitive to changes in its geometry. Conversely, the dif-ference with respect to the reference model is highest whenfriction coefficients are low for both faults (Figure 5). In thiscase, both faults are weak and likely to be activated. As longas frictional coefficients are low, a change toward a high dipangle on the MCT, which is mechanically unfavorable,makes the MHT comparatively much more efficient inaccommodating convergence. As a consequence, whencompared to the reference model, a 10° steeper MCT causesan ∼10 mm/yr transfer of slip to the MHT. Increasing thefriction on the MHT will tend to counterbalance this effectand progressively decrease the relative efficiency of theMHT, thereby leading to a diminishing difference withrespect to the reference model.[28] On the basis of slip rates on both faults and their

respective dip angles, we calculate the rock‐uplift rate abovethe MCT, due to the summed contributions of advection onthe MCT and on the ramp of the MHT located directlybelow (Figure 3). We observe that, even if fault slip rates aremodulated by the change in dip angle of the MCT, the actualrock uplift rate above the MCT is not significantly modifiedbetween the two models and the difference is actually lessthan 5% in the domain of friction coefficients that is com-patible with proposed kinematic scenarios (Figure 5d). Thetransfer of slip from one fault to another is also associatedwith correlated variations of the respective contributions torock uplift. These contributions globally compensate eachother, so that total rock uplift is roughly equal between thetwo configurations.

5.2. Position of the MCT

[29] Given the complex of faults that constitute the “MCTzone” [Searle and Godin, 2003; Hodges et al., 2004; Martinet al., 2005], we explore the impact of slip along a moreproximal or distal fault strand. In particular both the phys-iographic transition and faults identified by Hodges et al.[2004] are located south of traditionally mapped positionof the MCT. Hence it is relevant to assess the differences inmechanical response in differing spatial configurations. Forexample, shifting the MCT north or south by ±5 km withrespect to the Himalayan crest impacts the strain partitioningbetween the MHT and MCT, but with a lesser magnitudethan variations caused by changes in dip angle (Figure 6).

Deviations from the reference model are <3 mm/yr and, dueto this relatively narrow range of variations, additionaleffects related to modifications in the distributed strainaccommodation in the upper crust make the resulting totalresponse of the system more difficult to interpret.[30] Notably, shifting the MCT to the south by 5 km

brings it closer to the ramp of the MHT, such that the stressenvironment of the two faults becomes more similar,thereby reducing the contrast in behavior between them(Figures 6a and 6b). Such a situation diminishes themechanical advantage of transferring slip on the MCT whenfriction conditions on the MHT are less favorable. Forexample, for a strong MHT (m > 0.05) and a weak MCT(m < 0.15), which is a configuration where the systemwould tend to favor the transfer of slip toward the MCT(Figure 4), the proximity of the two faults causes themodeled slip to be lower on the MCT and larger on theMHT when compared to the reference model (Figures 6aand 6b). Similarly, when moving toward a weaker MHTand a stronger MCT, the slip on the MHT becomes pro-gressively lower than in the reference case (up to a ∼−2mm/yrdifference), whereas slip on the MCT becomes slightly faster(up to 0.25 mm/yr): changes that denote that the MCT iscomparatively more efficient at accommodating slip than inthe reference case. This relative efficiency is a directexpression of the diminished contrast between the two faults.The fact that contours for the 0 mm/yr difference in slip arenot coincident for the MCT and MHT indicates that addi-tional changes are predicted to occur in the relative contri-bution of distributed strain in the crust, as a response to themodification of the structure of the fault system. Globally,these results suggest that, in comparison to our referencemodel, an out‐of‐sequence activation of structures locatedprogressively farther south of the current MCT becomessuccessively more difficult. Conversely, a northward shift ofthe MCT has comparatively less impact on the slip on theMHT (Figure 6d). This shift tends to induce faster and slowerslip on the MCT and MHT, respectively, and thus corre-sponds to a situation where the contrast between the faults isamplified and the MCT appears to be slightly more efficientfrom amechanical point of view. The contrast is greater whenboth faults are weak, which represents a situation where theglobal fault system is more reactive to changes in the struc-tural configuration.

6. External Controls on Shortening Partitioning

[31] As illustrated by previous studies, variations in sur-face loads can modify the stress field and, thereby, have afirst‐order impact on the slip pattern of major faults [Hetzeland Hampel, 2005; Hampel and Hetzel, 2006; Hampelet al., 2007; Turpeinen et al., 2008; Hampel et al., 2009;Calais et al., 2010]. In our case, surface loads that canpotentially affect the MCT and the MHT are related both tothe temporal evolution of glaciation in the high range and tothe distribution of erosion in space and time. As previouslydescribed for the Himalayas of central Nepal [Pratt et al.,2002], denudation and, in particular, glacial erosion andice loading appear to display significant variations with timein that area.[32] Variations in surface loads modify both the normal

and shear stresses acting on the fault plane. In turn, these


9 of 19

stress changes modulate the Coulomb stress and promote orinhibit slip. The magnitude of normal and shear stresses is adirect function of the dip of the fault, such that fault dipbecomes a major factor controlling the effective sensitivityto surface‐load variations. For example, normal stress var-iations are likely to be significantly amplified on a gentlydipping fault. The impact of loads on stresses typically alsodepend on the depth along the fault plane, on the magnitudeof loading and unloading, and on the distance of the loadedarea from the fault.

6.1. Large‐Scale Erosion of the Topography

[33] For models in which no erosion is acting on thetopographic surface versus the reference model, the absenceof the modulating influence of erosion allows topography togrow quickly in areas of high rock uplift. The building of

high topography above major ramps, such as along theMFT, leads to a negative feedback, whereby the mass of therelief creates an increasing resistance that counter‐balancesthe tectonic forces and makes it harder to generate slip onthe corresponding structure (Figure 7). Such slip‐rate sen-sitivity to topographic loads is apparent on the MHT for lowfriction coefficients (Figure 7b): the development of thetopography in the Siwaliks will tend to inhibit the slip on theMHT, whereas rapid erosion of the weakly cemented fore-land strata in the MHT hanging wall [Lavé and Avouac,2000] enables slip. A similar phenomenon exists on theMCT for high friction coefficients on the MHT (Figure 7a),i.e., in a situation where the MCT is more likely toaccommodate a significant part of the convergence. On theother hand, when slip is slower on the MHT (for m < 0.02 onthe MHT and m > 0.15 on the MCT: Figure 7b), the slip rate

Figure 6. Influence of shifts in the north‐south position of the MCT on the slip partitioning between theMCT and the MHT. The MCT is shifted ±5 km with respect to its position in the reference model. Con-tour lines indicate the deviation in slip rates (mm/yr) from the reference model (Figure 4). (a) Differencein slip rate on the MCT with respect to the reference model when shifting the MCT 5 km southward(Figure 4a). (b) Difference in slip rate on the MHT with respect to the reference model when shifting theMCT 5 km southward (Figure 4b). (c) Difference in slip rate on the MCT with respect to the referencemodel when shifting the MCT 5 km northward (Figure 4a). (d) Difference in slip rate on the MHT withrespect to the reference model when shifting the MCT 5 km northward (Figure 4b). A southward shift ofthe MCT places it in a loading regime that is more similar to that of the MHT ramp, such that slip isinhibited on the MCT and transferred to the MHT.


10 of 19

on the MCT is not significantly different from the referencemodel (Figure 7a). Notably, these small changes indicatethat the resulting deficit in slip on the MHT is not trans-ferred to the MCT, but rather is accommodated as distrib-uted deformation (Figure 7c).

6.2. Time‐Varying Focused Erosion

[34] The models of the previous section were designed toprovide general insights on the large‐scale geodynamicalimpact of erosion, but do not explore the actual influence ofshorter wavelength variations in erosion patterns that arelikely to occur over several climatic cycles. As presented insection 2, erosion is expected to be highly variable with timein the high range due primarily to changes in glacial cov-erage and precipitation [Duncan et al., 1998; Pratt‐Sitaula,2005; Bookhagen et al., 2005a, 2005b; Godard et al., 2011].Here we present a series of simple sensitivity tests that aimat creating mechanical forcings similar to those resultingfrom fluctuations of glaciers in the High Himalaya.[35] We model the influence of focused glacial erosion in

the high range on fault slip (Figure 8). We use 0.01 and 0.15for the friction coefficients of the MHT and MCT, respec-tively, because these values are consistent with observed sliprates (Figure 4). A uniform erosion window is appliedacross the High Himalaya and southernmost Tibetan Plateau(between 120 and 150 km as indicated on Figures 1c and 3)in addition to the background erosion of the topography.This additional focused erosion is modulated through timeby a sine wave, with periods representing climatic cycles of10 and 40 ky, that continuously allows the glacial erosion tosweep from 0 up to the maximal defined value. This peri-odicity creates pulsed, active unloading in the high range thatis intended to explore the influence of episodic, rapid erosionby glaciers in an area which currently displays relatively lowinterglacial denudation rates [Gabet et al., 2008].[36] An increase in the intensity of glacial erosion induces

a progressive acceleration of slip on the MCT and the

development of larger finite slip than in the reference model(Figure 8b). For example, an erosion increase from ≤1 mm/yr(the modern observed interglacial rate: Gabet et al. [2008])to 5 mm/yr (a permissible glacial rate) causes about a 30%increase in the fault slip rate (Figure 8b). This acceleration isan expression of the progressive unclamping of the MCT dueto the unloading directly above the fault and the reduction ofnormal stresses. We note that time variations in the erosiveforcing, i.e., the 10‐ or 40‐ ky modulating sine wave, are nottemporally coupled with the slip on the MCT through time.This absence of a time‐dependent response suggests that thesystem is more sensitive to the long‐term cumulative effectsof the unloading, rather than to the short‐term variations inerosion. For the same erosion intensity, the response in termsof total slip on the MCT is apparently slightly higher with a10‐ky modulating period (Figure 8b), but this difference isactually due to a small difference in the cumulative erosionbetween the 10 and 40 ky forcings that appears during theonset of erosion in the simulation. In contrast to the MCT, theinfluence of the erosional unloading on the MHT slip is verylimited (Figure 8a). If the eroding area is shifted progres-sively to the north by increments of 10 km, more distalpositions for the glacial erosion with respect to the MCTinduce a decrease in the finite slip accommodated on thatfault (Figure 8c). This model result suggests a strong slipsensitivity to the position of the focused erosion with respectto the fault [Hampel et al., 2009].[37] Finally, we test the influence of combined spatial and

temporal variations in erosion. Our aim is to test the influ-ence of pulsed shifts of the locus of erosion such as thosethat might be associated with the alternation of glacialerosion in the high range during glacial periods with moreactive monsoon‐driven erosion on the southern flank of therange during interglacials. To study the response of themodel to this change in the site of maximum unloading, weconsider two sites of focused erosion: (1) one area between120 and 150 km (the high, glaciated Himalaya: Figure 1f)

Figure 7. Influence of erosion of the topography on the activity of the faults and the slip partitioningbetween the MCT and the MHT. Contours show the difference between runs without erosion and thereference model (Figure 4). (a) Difference in slip rate on the MCT with respect to the referencemodel (Figure 4a). (b) Difference in slip rate on the MHT with respect to the reference model (Figure 4b).(c) Difference between the applied velocity (20 mm/yr) and the combined slip rates on the MCT andMHT, with respect to the reference model (Figure 4c).


11 of 19

where we impose an additional erosion oscillating between0 and 2 mm/yr as in the previous models and (2) another onebetween 90 and 120 km (the monsoon‐dominated Hima-laya: Figure 1e) where we impose an additional erosionranging from 0 to 1 mm/yr. Both forcings are modulated bya sine wave with a 40‐ky period and are antiphased. Weobserve that, similar to our previous models, the oscillatingnature of the forcing is not visible in the slip history of theMCT (Figure 8b, dotted curve). We also note that the finite

slip is much higher than in the situation where only thenorthernmost “glacial” erosion is acting at the same maxi-mum rate of 2 mm/yr. This increase is a consequence of theadditional proximal unloading driven by monsoonal erosiondirectly above the MCT. These results support the conceptthat the MCT is mostly sensitive to the cumulative amountof erosion in the high range, rather than to its fine‐scaledistribution in space and time.

6.3. Influence of Ice Loads in the High Range

[38] In order to analyze the behavior of the system toloads that are more similar to those considered by previousstudies [Hetzel and Hampel, 2005; Turpeinen et al., 2008;Hampel et al., 2009], we impose pressure loads in the highrange between 120 and 150 km (Figures 1c and 3), that areequivalent to those imposed by 100 or 400 m of ice(Figures 4b and 9). This model scenario is designed to testthe influence of the weight of shrinking and expendingglaciers in the High Himalayas over the late Quaternary.We test two periods for the variations of the ice mass,corresponding to typical climatic variabilities of 10 and40 ky, and we observe that the loading slows down theMCT, but that the frequency of the oscillations is eitherimperceptible or very modest in both cases, even with themaximum load of 400 m of ice. This decrease in the MCTslip rate due to ice loads is consistent with the mechanicalexplanation proposed by earlier studies, where suchincreased loads induce an increase in the normal stressesacting on the fault plane and tend to clamp the fault.

Figure 9. Influence of an ice load applied on the topogra-phy between 120 and 150 km and modulated by a sine wavewith periods of 40 and 10 ky that makes the load fluctuatebetween 0 and the maximal imposed value (equivalent toeither 100 or 400 m of ice). The reference friction coeffi-cients are 0.01 and 0.15 on the MHT and MCT, respec-tively. The gray line is the reference model. The blackdashed and solid lines are models where loads equivalentto 100 and 400 m of ice are applied, respectively.

Figure 8. Influence of focused glacial erosion on slipactivity of the MHT and MCT. The reference friction coef-ficients are 0.01 and 0.15 on the MHT and MCT, respec-tively. Erosion is applied between 120 and 150 km(Figure 3) and modulated by a sine wave with a 10‐ or40‐ky period, that makes glacial erosion fluctuate between 0and the maximal defined value. (a) Influence on the MHTslip of different erosion intensities, ranging between 0 and5 mm/yr. (b) Influence of different erosion intensities on theMCT slip. The solid lines indicate the response of thesystem to focused erosion ranging from 0 to 5 mm/yr by1 mm/yr increments and modulated with 40‐ky periods.The dashed lines refer to erosion intensity of 2 and 5 mm/yr and modulated with 10‐ky periods. The dotted linecorresponds to a more complex erosion pattern wherefocused erosion shifts from south to north and back overperiods of 40 ky (see text for details), with a maximum valueof 1 mm/yr in the south (90–120 km) and 2 mm/yr in thenorth(120–150 km). (c) Decreased slip due to shifting theerosion zone northward by 10, 20, and 30 km with respect toits reference position. Erosion intensity is 3 mm/yr.


12 of 19

[39] We note two main differences with respect to theconclusions of some previous workers, such as Turpeinenet al. [2008]. First, total suppression of slip does not occurat the height of the loading phase. As explained above,continued slip is a consequence of the much higher tectonicloading rates in our model. These rates prevail over externalparameters in controlling the slip activity of the faults.Second, the slip deficit is not recovered during the unload-ing because slip has been transmitted to MHT, such that nostress accumulation results from the decrease in the slip rateof the MCT. In this sense, the ability to partition deforma-tion between the MCT and MHT decreases the sensitivity ofthe MCT to surface loads.[40] The 400‐m value for the ice load is generally above

the estimates for ice coverage over in the high range ofcentral Nepal during the local Last Glacial Maximum[Pratt‐Sitaula, 2005]. This upper bound, therefore, suggeststhat ice loading and unloading alone can not be a cause ofsignificant modulation of the strain partitioning, althoughthis conclusion is dependent on the values we use for therheological parameters and reference friction coefficients.We predict that lower friction coefficients will induce agreater sensitivity of the system to external loads.

6.4. Time‐Varying Friction Coefficients

[41] The MCT is a zone of intense fluid circulation, with ahydrothermal system closely connected to surface hydrol-ogy [Derry et al., 2009], and it has been shown in otherregions that meteoritic water could penetrate deep in the

upper crust [Craw et al., 1994]. We speculate that variationsin water availability at the surface that are likely to berelated to changing climatic conditions over time may betransmitted to the shallow parts of the MCT. This fluid fluxwould induce a change in physical conditions in the faultzone in terms of lubrication and pore pressure that couldimpact the effective friction. Although we acknowledge thatthe existence of such an effect can not be readily validatedwith the currently available geological and geophysical data,we nevertheless test several scenarios of friction coefficientvariations for the MCT (Figure 10). Our aim here is toprovide further insights on the mechanical sensitivity of thesystem to variations in its boundary conditions. Thus, wetest the response of the orogen to oscillations in the frictioncoefficient of the MCT around its reference value withdifferent periods that are typical of climatic times scales: 10and 40 ky (Figure 10).[42] For these different frequencies, small variations of

the friction coefficient are directly expressed as significantslowing or acceleration of the slip rate on the MCT. The40‐ky oscillations with 5% changes in frictional coeffi-cients induce about 50% changes in predicted slip rates(Figure 10a). In the case of the higher frequency oscilla-tions (10 ky), the increase of the friction coefficient by only5% results in a total suppression of slip, whereas a com-parable reduction in the coefficient more than doubles thefault slip rate (Figure 10b). We also note that the 10‐kyoscillations result in a finite slip at the end of the simulationthat is significantly higher than in the reference model. Thisincrease is a consequence of the temporary suppression ofslip of the MCT that induces systematic overshooting whenthe fault is reactivated. This overshot only appears for thehigh frequency oscillations, which suggests that it resultsfrom the short‐term accumulation of elastic energy in thedeforming wedge, when the MCT is locked, that can not beefficiently dissipated by increasing the slip on the MHT.These simulations illustrate the idea that relatively smallvariations in the friction conditions on the fault plane couldresult in pronounced and geodynamically significant effectsin terms of slip rate on crustal‐scale faults at Milankovitchtimescales.

7. Discussion

7.1. Implications for Himalayan Tectonics

[43] Our study presents a detailed and systematic analysisof the mechanical behavior of the MCT/MHT system in theHimalayas (Figure 11). Whereas some of the responses ofthe models to the various changes in boundary conditionsthat we have explored here can be intuited from simpleconsiderations on the mechanical efficiency of the faults andminimum work principle, the systematic exploration ofthese responses allows derivation of useful quantitativeinformation on the modalities of slip partitioning.[44] Our models suggest that variations of the friction

coefficients on either fault have major implications for thelarge‐scale deformation pattern of the Himalayas. Strongfrictional contrasts can induce highly contrasting kinemat-ics, wherein either fault can act as the dominant structure forthe accommodation of overall convergence. More im-portantly, we note that measured rates of Holocene slipimpose strict bounds on the values of the friction coeffi-

Figure 10. Slip variations on the MCT due to varying thefriction coefficient for it. The reference friction coefficientsare 0.01 and 0.15 on the MHT and MCT, respectively. Thesolid gray curve is the reference model (no variations of thecoefficient with time). The dashed and solid gray lines arecases where the friction coefficient is modulated with a sinewave by ±2% and ±5% around the reference value, respec-tively. (a) 40‐ky period for the modulation. (b) 10‐ky periodfor the modulation.


13 of 19

cients. In order to reproduce the situation where ∼20 mm/yrof slip is accommodated on the MHT and little or no dis-placement occurs on the MCT [Lavé and Avouac, 2001;Herman et al., 2010a], with our reference model geometryand rheological parameters, we have to impose frictioncoefficients <0.02 and >0.1 on the MHT and MCT,respectively. Notably, this range in the friction‐coefficientspace (Figure 4) is characterized by high slip‐rate gradients,such that slight modifications of the friction conditions oneither the MCT or the MHT can result in significant mod-ification of the slip partitioning between the faults.[45] We tested the influence of several parameters on the

response of the mechanical system, in terms of partitioningbetween the two faults. The key controlling parameterappears to be the dip angle of the MCT, because it directlycontrols the mechanical and energy efficiency of slipaccommodation on that structure: a steeper MCT (30°versus the 20° of the reference model) is much less likelyto accrue significant slip (Figure 5). The overriding controlthat is exerted by the MCT’s steepness has notable im-plications for the understanding of along‐strike variationsof the present and Quaternary kinematics in the CentralHimalayas. A consequence of our modeling is that a steepMCT is going to favor the kinematic scenarios advocatedby Robert et al. [2009] or Herman et al. [2010a] in whichout‐of‐sequence thrusting on the MCT is not significant atthe range scale. Similarly, if the MCT is substantiallysteeper along a trans‐Himalayan transect through Kath-mandu [Pearson and DeCelles, 2005] than in a transectalong the Marsyandi 75 km farther west, our modelingpredicts that the MCT might be inactive in the easterntransect while it is actively slipping in the west. Such aslip pattern would introduce a lateral gradient in shorten-ing along the MCT that should be accommodated pro-gressively along strike inasmuch as no discrete transformstructure has been documented along this portion of therange.

[46] Following this point, it is noteworthy that a numberof major differences exist in the along‐strike structural andgeomorphic properties of the Himalayas of central Nepal.One striking feature is the Kathmandu klippe, where HigherHimalayan material up to 6–10 km thick has been overthrustacross the Lesser Himalaya for up to ∼50 km south of thephysiographic transition (PT2 [Wobus et al. 2003]). Nota-bly, the PT2 is nearly continuous along strike in centralNepal and if we assume that it is a morphologic expressionof thrusting activity at the base of the Higher Himalaya[Wobus et al., 2003], then its presence does not support theidea of large lateral variations in rates of out‐of‐sequencethrusting. Moreover, as suggested by Bollinger et al. [2004a],lateral variations exist in the degree of development of theLesser Himalayan duplex system due to the along‐strikesuccession of Higher Himalayan klippes or Lesser Hima-layan windows. We can speculate that the magnitude of thestacking of nappes of Lesser Himalayan origin in thefootwall of the MCT has an influence on the present dip ofthe MCT and may result in significant along‐strike dipvariations. As suggested by our modeling (Figure 5), suchdip variations are expected to exert a first‐order control onthe behavior of the orogen and on the pattern of slippartitioning between the faults. Another noteworthy con-sequence of slip transfer between the two faults dependingon the structural configuration is that, in our modeling, therock uplift above the MCT is not very sensitive to thepartitioning pattern, due to negatively correlated slip var-iations between the MCT and the underlying ramp of theMHT (Figure 5d). We also note that the idea of significantQuaternary slip of the MCT or nearby underlying faultshas been sparked by data collected west of the Kathmanduarea. Some of the initial results are from the Buri Gandakicatchment [Wobus et al., 2003, 2005] and are close toareas where the recent analysis of other data sets andthermo‐kinematic modeling tend to argue against a sig-nificant out‐off‐sequence motion of a structure locatedalong the PT2 or the MCT [Robert et al., 2009; Herman

Figure 11. Summary of the main factors that influence the partitioning of deformation across the MHT/MCT fault system in terms of accelerating or decelerating the slip rate on those faults, namely, the dipangle of the MCT (Figure 5), large scale erosion across the range (Figure 7) and the focused unload-ing in the high range (Figure 8).


14 of 19

et al., 2010a]. In contrast, farther west in the Marsyandicatchment, a localized offset in cooling ages exists (Figure 2)and has been used to propose reactivation of the MCT there[Huntington et al., 2006; Blythe et al., 2007], whereas arecent study by Robert et al. [2011] suggests that out‐of‐sequence reactivation is not required to explain the distri-bution of thermochronological ages in this region of theMarsyandi and Kali Gandaki. Clearly more data are neededto document the patterns of exhumation in the HimalayasNepal and discriminate among the different models, but wepropose that an understanding of the kinematic variabilityalong the Himalayan arc should focus on variations in thedip of the structures as one of the primary explanatoryparameters.

7.2. Influence of Surface Loads

[47] Several studies have already used generic models toinvestigate the influence of ice or erosional loading andunloading in modulating the slip of different types of faults[Hampel and Hetzel, 2006; Turpeinen et al., 2008; Maniatiset al., 2009]. Those studies have shown that, in the caseof far‐field extension or convergence rates of the order of2 mm/yr, variations in surface loads could modulate theactivity of the faults embedded in the model. One of themajor differences between our modeling and those studiesis that we impose a shortening velocity that mimics moderngeodetic rates across the Himalaya and is an order ofmagnitude higher on the model domain. In this situation offaster and more intense tectonic processes, it might beexpected that the relative impact of similar surface loads onfault activity is likely to be less pronounced and partlyhidden behind the background tectonic signal. In ourmodels, fault slip is primarily driven by the imposed hor-izontal stresses and the possible range of variations asso-ciated with modulation of vertical stresses is relativelynarrow. We also note that the ice loads imposed in theseprevious studies are equivalent to several hundreds metersof ice, whereas we doubt that spatially averaged ice loadswould exceed 100 m in the glaciated Himalaya due to thetypical hillslope steepness (>20°). Thus, our study suggeststhat slip‐rate modulation by glacial or erosional loads islikely to be less efficient in environments characterized byhigh rates of tectonic loading. Nonetheless, the fission trackdata [Burbank et al., 2003; Blythe et al., 2007] implyaverage erosion rates in the glacial‐periglacial Himalayanrealm have been significantly faster than those observedtoday. Our analysis suggests that increasing the erosion rateto 4 mm/yr (rates consistent with the fission track ages)from today’s rate of 1 mm/yr would increase the magnitudeof slip on the MCT by about 30% (Figure 8). However,even with such a significant increase in slip rate, erosionalunloading alone appears insufficient to explain the fulloffset in cooling ages observed along the Marsyandi river(Figure 2).[48] A clear sensitivity of fault zones to surface loads at

shorter time‐scales has also been highlighted by severalprevious workers [e.g., Luttrell et al., 2007]. In particular,two recent studies have demonstrated that the tectonicregime of the Himalayas of central Nepal could displaysignificant variations at annual timescales as a response toseasonal variations in surface loads. Using continuous GPStime series, Bettinelli et al. [2008] have shown that, across

the range, the strain pattern with respect to a long‐term trenddisplayed clear seasonal fluctuations, which they attributedto the cyclic loading of the Gangetic plain by monsoonrainwater. Similarly, from the investigation of micro‐seismicity databases, Bollinger et al. [2007] demonstratedthe existence of a statistically significant variation in theoccurrence of Himalayan earthquakes between the winterand the summer. They also interpret the relative decrease ofseismicity in the summer as a consequence of monsoonloads in the Himalayan foreland. Whereas these studies bothdeal with surface‐loading processes at much shorter time-scales than the ones we investigate here and activate dif-ferent physical mechanisms, such as perturbations of theelastic interseismic strain field, they clearly demonstrate anelevated sensitivity of the Himalayan tectonic behavior toexternal perturbations. The processes presented by those twostudies are by nature oscillatory and of low magnitude, butthey ultimately tend to support the idea that infinitesimal andunidirectional unloading on an annual basis can result in afinite alteration of the tectonic deformation if it accumulatesover the long term. This summation may lead to an excess ordeficit in slip on major faults, such as that which we modelfor the MCT as a consequence of focused unloading in theHigh Himalaya (Figure 8).

7.3. Mechanical Behavior at Scales of a Fault orOrogenic Wedge

[49] In a similar modeling approach, Cattin and Avouac[2000] proposed that the friction coefficients could be upto 0.6 on the ramp of the MHT and up to 0.13 on the flatportion of this fault. These values are much higher thanthose we infer in our study for the Holocene kinematicconfiguration. This seeming inconsistency is a direct con-sequence of the possibility that, in our model, the mechan-ical system can transfer convergence to the MCT, if slip onthe MHT is not energetically favorable. Such partitioningwas not possible in Cattin and Avouac’s model and appar-ently allowed higher friction coefficients on the MHT thanare viable in our model. We note that the thermo‐kinematicmodeling approach proposed by Herman et al. [2010a] thatconsiders friction‐dependent shear heating on the mainfaults also leads to the conclusion that the friction coefficientneeds to be low on the MHT.[50] The range of the friction coefficients inferred for the

MHT suggests a relatively weak fault (Figure 4), but isconsistent with constraints derived for major continentalcrustal‐scale faults in Tibet [He and Chéry, 2008], along theSan Andreas fault zone [Lachenbruch and Sass, 1980;Zoback and Beroza, 1993; Faulkner and Rutter, 2001;Parsons, 2002] or in subduction zones [Magee and Zoback,1993]. The contrast in friction coefficients between theMHT and MCT might be a consequence of the high avail-ability of fluids along the MHT, derived from the meta-morphism of sediments of the subducting Indian plate, asinferred from geophysical imaging [Nabelek et al., 2009]. Incontrast, whereas hydrothermal convection is intense in theMCT zone in the near surface [Derry et al., 2009], its at‐depth extension is speculative and is not observable withavailable geophysical data. Climatically driven fluid vari-ability may only affect a rather superficial portion of thefault, may be buffered by a vigorous groundwater system, ormay have no large‐scale impact on fault friction and


15 of 19

behavior (Figure 10). However, the fact that very modestvariations in friction coefficient can significantly modulateslip on the fault without being overridden by the backgroundrates of tectonic loading suggests that intrinsic fault zoneproperties might need to be further considered when dis-cussing the variability of fault slip rates over time.[51] From a broader perspective, the causes of localization

of deformation in active orogenic systems is one of the mostimportant issues for our understanding of orogenic dynam-ics. Such a problem is often formulated in terms of energeticefficiency on the different faults that may accommodatedeformation [Cubas et al., 2008], and the system shouldtend to favor deformation scenarios that minimize the inte-grated amount of work required to cope with the far‐fieldshortening, either by developing new structures, outward orinward, or by reactivating pre‐existing faults. Our study isconsistent with this minimum work principle, but alsoillustrates the idea that, in the case of a geometricallycomplex orogen, slight modifications of internal propertiesor boundary conditions can lead to substantial shifts inthe deformation pattern and to evolutionary sequences thatare either sustained or transient. Previous studies havehighlighted some specific modes of evolution of orogenicwedges that develop in a changing climate [Stolar et al.,2006; Whipple and Meade, 2006; Whipple, 2009], wherethe outward expansion of the range or reduction in itswidth are directly connected to changes in climaticallycontrolled erosional efficiency. Most of these studies usedcritical wedge formulations [Dahlen, 1990] that do notallow assessment of how the strain response to the cli-matic forcing is distributed on discrete faults. On theother hand, our study introduces pre‐defined fault zonesas contact interfaces in the model (Figure 3), but does notallow for the development of new shortening zones duringthe simulation. Analog modeling approaches that allowfor both localized deformation on fault‐like structures, aswell as dynamic development and abandonment of thesestructures, have highlighted a complex temporal evolutionof orogenic wedge behavior [Konstantinovskaia andMalavieille, 2005; Hoth et al., 2007], with complex cy-cles of fault activity and distribution of deformationacross the wedge at climatic time‐scales. The sensitivityof strain partitioning inside the wedge to variations inerosion that is illustrated by our study (Figure 8) andothers [Maniatis et al., 2009], suggests that the internalclock of Hoth et al. [2007] and forcing by fluctuatingclimate, especially when acting at similar time‐scales, areprobably intimately intertwined in dictating the actualtectonic evolution and changes in the distribution ofdeformation of active orogens.

7.4. Limitations and Future Research Directions

[52] We have used a simplified geometry for the contactinterfaces intended as representations of the MCT and MHTin our numerical model, where they are modeled as acombination of straight‐line segments. Although severalarguments can be used to posit a flat/ramp geometry for theMHT in central Nepal [Schelling and Arita, 1991; Pandeyet al., 1995; Lavé and Avouac, 2001], the evolution of theMCT dip at depth remains unknown, and whether thisstructure becomes progressively gentler as it merges with amid‐crustal basal thrust is a question that still needs to be

answered, possibly by high‐resolution geophysical imagery[Schulte‐Pelkum et al., 2005; Nabelek et al., 2009]. Nev-ertheless, we anticipate that variations at depth of the dipangle of the MCT, with respect to the reference dip of theupper part of the fault, will have similar effects on slippartitioning to that which was observed for the whole MCTin this study (Figure 5). Steepening at depth (i.e., formationof a ramp) will lead to slip transfer toward the MHT, butwould probably be less pronounced because it will affectonly part of the fault.[53] Independently of the degree of complexity we

assume for the faults, our sensitivity tests on the dip angle ofthe MCT (Figure 5) demonstrate that structural variationscan have a first‐order impact on the kinematics of the range.On the basis of those results, we argue that the fine‐scaledocumentation of the along‐strike variations in the geometryof both the MHT and MCT is probably one of the mostimportant elements for understanding the kinematics of theHimalayan range at 10–100 ky time‐scales.[54] Our study was initially spurred by the intriguing

temporal variations of erosion in the high range [Burbanket al., 2003; Gabet et al., 2008] and the sharp offset incooling ages across the MCT zone [Huntington et al.,2006; Blythe et al., 2007] in the Marsyandi catchment.We acknowledge, however, that no unambiguous evidenceexists for a significant Holocene activity of the MCT inthat area, even if several data sets in adjacent areas pointtoward recent tectonic activity focused in the physiographictransition [Wobus et al., 2003; Hodges et al., 2004; Wobuset al., 2005]. Decisive field evidence still needs to beprovided to assess the existence and, if present, the mag-nitude of recent slip on the MCT. We note that recentdevelopments in low‐temperature thermochronology havemade available new dating techniques with very low clo-sure temperatures [Herman et al., 2010b], that would havethe potential to detect very recent and relatively smallamplitude variations in exhumation across the MCT zone.Detailed geomorphic mapping in this steep, rapidly erod-ing, and monsoon‐drenched southern flank of the HighHimalayas is difficult but high‐resolution analysis ofstream profiles [Craddock et al., 2007] and CRN inven-tories in stream sediments [Wobus et al., 2005] also havethe potential to yield other pieces of evidence.

8. Conclusions

[55] Strain partitioning between major shortening struc-tures is a key feature of deformation within orogenicwedges. Whereas increasingly detailed documentation offault slip rates can provide an overall kinematic picture ofthis process, the mechanical aspects of partitioning are stillpoorly resolved. In the Himalaya of central Nepal, theproblem of slip partitioning between thrust faults, such asthe Main Himalayan Thrust and the Main Central Thrust, isstill actively debated. Whereas this problem has been pre-viously explored using thermo‐kinematic approaches[Brewer and Burbank, 2006; Whipp et al., 2007; Robertet al., 2009; Herman et al., 2010a], no clear mechanicalframework has been proposed that could provide a physicalunderstanding of the partitioning between those two faults.Our study is an attempt to deliver this kind of perspective onthe behavior of the range and should help the joint analysis


16 of 19

of the diversity of structural styles, geomorphologic con-trols, and kinematic regimes that are observed along theHimalayan belt.[56] Various modes of interactions between tectonics and

surface processes in active mountain ranges have beenproposed over the last 20 years. Our study suggests theexistence of an intriguing sensitivity of the Himalayanwedge to surface loads. In the context of background sliprates imposed by the far‐field tectonic loading, our modelsindicate that reasonable unloading configurations couldsignificantly modulate slip rates on major faults. Whereassuch responses are dependent on the assumptions made forcrustal rheology and fault strength, our results suggest theexistence of significant transient adjustments at climatictime‐scales (10–100 ky) such that the idea of tectonic steadystate in an active mountain range should be used withcaution and in a time frame that exceeds the scale of short‐term variability.

[57] Acknowledgments. This research was funded by a Marie CurieFellowship (Geocycl‐219662) to V.G. and by NSF (EAR 0819874) andNASA (NNX08AG05G) grants to D.W.B. The authors acknowledge thesupport of the Institute of Crustal Studies and, in particular, Chen Ji foraccess to computational facilities. We thank B. Bookhagen, G. B. Fisher,and J.‐P. Avouac for helpful discussions during the development of thisproject. Constructive comments by Frédéric Herman, Peter van der Beek,Associate Editor W. P. Schellart, and one anonymous reviewer helped toimprove this manuscript.

ReferencesArmijo, R., P. Tapponnier, J. L. Mercier, and H. Tong‐Lin (1986), Quater-nary extension in southern Tibet: Field observations and tectonic implica-t ions , J . Geophys . Res . , 91 , 13 ,803–13,872 , do i :10 .1029/JB091iB14p13803.

Armstrong, R., B. Raup, S. Khalsa, R. Barry, J. Kargel, C. Helm, andH. Kieffer (2005), Glims glacier database, Tech. rep., Natl. Snowand Ice Data Cent., Boulder, Colo.

Avouac, J.‐P. (2003), Mountain building, erosion, and the seismic cycle inthe Nepal Himalaya, Adv. Geophys., 46, 1–80.

Avouac, J.‐P. (2007), Dynamic processes in extensional and compressionalsettings–Mountain building: From earthquakes to geological deforma-tion, in Treatise of Geophysics, vol. 6, Crust and Lithosphere Dynamics,edited by A. B. Watts, pp. 377–439, Elsevier, Boston, doi:10.1016/B978-044452748-6.00112-7.

Avouac, J.‐P., and E. B. Burov (1996), Erosion as a driving mechanism ofintracontinental mountain growth, J. Geophys. Res., 101, 17,747–17,770,doi:10.1029/96JB01344.

Beaumont, C., P. Fullsack, and J. Hamilton (1992), Erosional control ofactive compressional orogens, in Thrust Tectonics, vol. 99, pp. 1–18,Chapman and Hall, London.

Beaumont, C., R. A. Jamieson, M. H. Nguyen, and B. Lee (2001), Hima-layan tectonics explained by extrusion of a low‐viscosity crustal channelcoupled to focused surface denudation, Nature, 414, 738–742,doi:10.1038/414738a.

Berger, A. L., et al. (2008), Quaternary tectonic response to intensifiedglacial erosion in an orogenic wedge, Nat. Geosci., 1, 793–799,doi:10.1038/ngeo334.

Bettinelli, P., J. Avouac, M. Flouzat, F. Jouanne, L. Bollinger, P. Willis,and G. Chitrakar (2006), Plate motion of India and interseismic strainin the Nepal Himalaya from GPS and DORIS measurements, J. Geod.,80(8), 567–589, doi:10.1007/s00190-006-0030-3.

Bettinelli, P., J. Avouac, M. Flouzat, L. Bollinger, G. Ramillien, S. Rajaure,and S. Sapkota (2008), Seasonal variations of seismicity and geodeticstrain in the Himalaya induced by surface hydrology, Earth Planet.Sci. Lett., 266, 332–344, doi:10.1016/j.epsl.2007.11.021.

Bilham, R., K. M. Larson, and J. T. Freymueller (1997), GPS measure-ments of present‐day convergence across the Nepal Himalaya, Nature,386, 61–64, doi:10.1038/386061a0.

Blythe, A., D. Burbank, A. Carter, K. Schmidt, and J. Putkonen (2007),Plio‐Quaternary exhumation history of the central Nepalese Himalaya:1. Apatite and zircon fission track and apatite [U‐Th]/He analyses,Tectonics, 26, TC3002, doi:10.1029/2006TC001990.

Bollinger, L., J. P. Avouac, O. Beyssac, E. J. Catlos, T. M. Harrison,M. Grove, B. Goffé, and S. Sapkota (2004a), Thermal structure andexhumation history of the Lesser Himalaya in central Nepal, Tectonics, 23,TC5015, doi:10.1029/2003TC001564.

Bollinger, L., J. P. Avouac, R. Cattin, and M. R. Pandey (2004b), Stressbuildup in the Himalaya, J. Geophys. Res., 109, B11405, doi:10.1029/2003JB002911.

Bollinger, L., F. Perrier, J. P. Avouac, S. Sapkota, U. Gautam, and D. R.Tiwari (2007), Seasonal modulation of seismicity in the Himalaya ofNepal, Geophys. Res. Lett., 34, L08304, doi:10.1029/2006GL029192.

Bookhagen, B., and D. W. Burbank (2006), Topography, relief, andTRMM‐derived rainfall variations along the Himalaya, Geophys. Res.Lett., 33, L08405, doi:10.1029/2006GL026037.

Bookhagen, B., R. C. Thiede, and M. R. Strecker (2005a), Late Quaternaryintensified monsoon phases control landscape evolution in the northwestHimalaya, Geology, 33(2), 149–152.

Bookhagen, B., R. C. Thiede, and M. R. Strecker (2005b), Abnormal mon-soon years and their control on erosion and sediment flux in the high, aridnorthwest Himalaya, Earth Planet. Sci. Lett., 231, 131–146.

Brewer, I. D., and D. W. Burbank (2006), Thermal and kinematic modelingof bedrock and detrital cooling ages in the central Himalaya, J. Geophys.Res., 111, B09409, doi:10.1029/2004JB003304.

Brewer, I. D., D. W. Burbank, and K. V. Hodges (2006), Downstreamdevelopment of a detrital cooling‐age signal: Insights from 40Ar/39Armuscovite thermochronology in the Nepalese Himalaya, in Tectonics,Climate, and Landscape Evolution, edited by S. D. Willett et al., Spec.Pap. Geol. Soc. Am., 398, 321–338.

Burbank, D. W., A. E. Blythe, J. Putkonen, B. Pratt‐Sitaula, E. Gabet,M. Oskin, A. Barros, and T. P. Ojha (2003), Decoupling of erosionand precipitation in the Himalayas, Nature , 426 , 652–655,doi:10.1038/nature02187.

Calais, E., A. M. Freed, R. Van Arsdale, and S. Stein (2010), Triggeringof New Madrid seismicity by late‐Pleistocene erosion, Nature, 466,608–611, doi:10.1038/nature09258.

Carter, N. L., and M. C. Tsenn (1987), Flow properties of continentallithosphere, Tectonophysics, 136(1–2), 27–63, doi:10.1016/0040-1951(87)90333-7.

Catlos, E., T. M. Harrison, M. Kohn, M. Grove, M. Ryerson, C. E.Manning, and B. N. Upreti (2001), Geochronologic and thermobaro-metric constraints on the evolution of the Main Central Thrust, centralNepal Himalaya, J. Geophys. Res., 106, 16,177–16,204, doi:10.1029/2000JB900375.

Cattin, R., and J. P. Avouac (2000), Modeling mountain building and theseismic cycle in the Himalaya of Nepal, J. Geophys. Res., 105,13,389–13,407, doi:10.1029/2000JB900032.

Cattin, R., G. Martelet, P. Henry, J.‐P. Avouac, M. Diament, and T. R.Shakya (2001), Gravity anomalies, crustal structure and thermo‐mechanicalsupport of the Himalaya of central Nepal,Geophys. J. Int., 147(2), 381–392,doi:10.1046/j.0956-540x.2001.01541.x.

Craddock, W. H., D. W. Burbank, B. Bookhagen, and E. J. Gabet (2007),Bedrock channel geometry along an orographic rainfall gradient in theupper Marsyandi River valley in central Nepal, J. Geophys. Res., 112,F03007, doi:10.1029/2006JF000589.

Craw, D., P. O. Koons, D. Winslow, C. P. Chamberlain, and P. Zeitler(1994), Boiling fluids in a region of rapid uplift, Nanga Parbat massif,Pakistan, Earth Planet. Sci. Lett., 128, 169–182, doi:10.1016/0012-821X(94)90143-0.

Cubas, N., Y. M. Leroy, and B. Maillot (2008), Prediction of thrusting se-quences in accretionary wedges, J. Geophys. Res., 113, B12412,doi:10.1029/2008JB005717.

Dahlen, F. (1990), Critical taper model of fold‐and‐thrust belts and accre-tionary wedges, Annu. Rev. Earth Planet. Sci., 18(1), 55–99.

DeCelles, P., D. Robinson, J. Quade, T. Ojha, C. Garzione, P. Copeland,and B. Upreti (2001), Stratigraphy, structure, and tectonic evolution ofthe Himalayan fold‐thrust belt in western Nepal, Tectonics, 20(4),487–509, doi:10.1029/2000TC001226.

Derry, L. A., M. J. Evans, R. Darling, and C. France‐Lanord (2009),Hydrothermal heat flow near the Main Central Thrust, central NepalHimalaya, Earth Planet. Sci. Lett., 286, 101–109, doi:10.1016/j.epsl.2009.06.036.

Duncan, C. C., A. J. Klein, J. G. Masek, and B. L. Isacks (1998), Compar-ison of Late Pleistocene and modern glacier extents in central Nepalbased on digital elevation data and satellite imagery, Quat. Res., 49(3),241–254, doi:10.1006/qres.1998.1958.

Faulkner, D. R., and E. H. Rutter (2001), Can the maintenance of overpres-sured fluids in large strike‐slip fault zones explain their apparent weak-ness?, Geology, 29(6), 503–506.

Friedrich, A. M., B. P. Wernicke, N. A. Niemi, R. A. Bennett, and J. L.Davis (2003), Comparison of geodetic and geologic data from the


17 of 19

Wasatch region, Utah, and implications for the spectral character ofEarth deformation at periods of 10 to 10 million years, J. Geophys.Res., 108(B4), 2199, doi:10.1029/2001JB000682.

Fuller, C. W., S. D. Willett, and M. T. Brandon (2006), Formation of fore-arc basins and their influence on subduction zone earthquakes, Geology,34(2), 65–68, doi:10.1130/G21828.1.

Gabet, E. J., B. A. Pratt‐Sitaula, and D. W. Burbank (2004a), Climaticcontrols on hillslope angle and relief in the Himalayas, Geology, 32(7),629–632.

Gabet, E. J., D. W. Burbank, J. K. Putkonen, B. A. Pratt‐Sitaula, andT. Ojha (2004b), Rainfall thresholds for landsliding in the Himalayasof Nepal, Geomorphology, 63(3–4), 131–143, doi:10.1016/j.geo-morph.2004.03.011.

Gabet, E. J., D. W. Burbank, B. Pratt‐Sitaula, J. Putkonen, andB. Bookhagen (2008), Modern erosion rates in the High Himalayas ofNepal, Earth Planet. Sci. Lett. , 267, 482–494, doi:10.1016/j.epsl.2007.11.059.

Gabet, E. J., D. Wolff‐Boenisch, H. Langner, D. W. Burbank, andJ. Putkonen (2010), Geomorphic and climatic controls on chemical weath-ering in the High Himalayas of Nepal, Geomorphology, 122(1–2),205–210, doi:10.1016/j.geomorph.2010.06.016.

Gansser, A. (1964), Geology of the Himalayas, vol. 289, Interscience,London.

Garzanti, E., G. Vezzoli, S. Andò, J. Lavé, M. Attal, C. France‐Lanord, andP. DeCelles (2007), Quantifying sand provenance and erosion (MarsyandiRiver, Nepal Himalaya), Earth Planet. Sci. Lett., 258, 500–515,doi:10.1016/j.epsl.2007.04.010.

Gayer, E., J. Lavé, R. Pik, and C. France‐Lanord (2006), Monsoonal forc-ing of Holocene glacier fluctuations in Ganesh Himal (central Nepal)constrained by cosmogenic 3He exposure ages of garnets, Earth Planet.Sci. Lett., 252, 275–288, doi:10.1016/j.epsl.2006.09.040.

Godard, V., R. Cattin, and J. Lavé (2004), Numerical modeling ofmountain building: Interplay between erosion law and crustal rheol-ogy, Geophys. Res. Lett., 31, L23607, doi:10.1029/2004GL021006.

Godard, V., J. Lavé, and R. Cattin (2006), Numerical modelling of erosionprocesses in the Himalayas of Nepal: Effects of spatial variations of rockstrength and precipitation, in Analogue and Numerical Modelling ofCrustal‐Scale Processes, edited by S. J. H. Buiter and G. Schreurs, Geol.Soc. Spec. Publ., 253, 341–358, doi:10.1144/GSL.SP.2006.253.01.18.

Godard, V., R. Cattin, and J. Lavé (2009), Erosional control on the dynam-ics of low‐convergence rate continental plateau margins, Geophys. J. Int.,179(2), 763–777, doi:10.1111/j.1365-246X.2009.04324.x.

Godard, V., D. W. Burbank, D. L. Bourlès, B. Bookhagen, R. Braucher,and G. B. Fisher (2011), Constraining catchment‐scale glacial erosionusing a global inversion of detrital CRN data in the Himalayas of centralNepal, in Geophys. Res. Abstr., 13, EGU2011‐1447.

Goodbred, S. L., and S. A. Kuehl (1999), Holocene and modern sedimentbudgets for the Ganges‐Brahmaputra river system: Evidence for high-stand dispersal to flood‐plain, shelf, and deep‐sea depocenters, Geology,27(6), 559–562.

Hampel, A., and R. Hetzel (2006), Response of normal faults to glacial‐interglacial fluctuations of ice and water masses on Earth’s surface,J. Geophys. Res., 111, B06406, doi:10.1029/2005JB004124.

Hampel, A., R. Hetzel, and A. L. Densmore (2007), Postglacial slip‐rateincrease on the Teton normal fault, northern Basin and Range Province,caused by melting of the Yellowstone ice cap and deglaciation of theTeton Range?, Geology, 35(12), 1107–1110.

Hampel, A., R. Hetzel, G. Maniatis, and T. Karow (2009), Three‐dimensionalnumerical modeling of slip rate variations on normal and thrust fault arraysduring ice cap growth and melting, J. Geophys. Res., 114, B08406,doi:10.1029/2008JB006113.

Harper, J. T., and N. F. Humphrey (2003), High altitude Himalayan climateinferred from glacial ice flux, Geophys. Res. Lett., 30(14), 1764,doi:10.1029/2003GL017329.

Harrison, T.M.,M. Grove, O.M. Lovera, and E. J. Catlos (1998), Amodel forthe origin of Himalayan anatexis and inverted metamorphism, J. Geophys.Res., 103, 27,017–27,032.

Hassani, R., D. Jongmans, and J. Chéry (1997), Study of plate deformationand stress in subduction processes using two‐dimensional numericalmodels, J. Geophys. Res., 102, 17,951–17,966, doi:10.1029/97JB01354.

He, J., and J. Chéry (2008), Slip rates of the Altyn Tagh, Kunlun and Kar-akorum faults (Tibet) from 3D mechanical modeling, Earth Planet. Sci.Lett., 274, 50–58, doi:10.1016/j.epsl.2008.06.049.

Henry, P., X. Le Pichon, and B. Goffé (1997), Kinematic, thermal and pet-rological model of the Himalayas: Constraints related to metamorphismwithin the underthrust Indian crust and topographic elevation, Tectono-physics, 273(1–2), 31–56, doi:10.1016/S0040-1951(96)00287-9.

Herman, F., et al. (2010a), Exhumation, crustal deformation, and thermalstructure of the Nepal Himalaya derived from the inversion of thermo-

chronological and thermobarometric data and modeling of the topogra-phy, J. Geophys. Res., 115, B06407, doi:10.1029/2008JB006126.

Herman, F., E. J. Rhodes, J. Braun, and L. Heiniger (2010b), Uniform ero-sion rates and relief amplitude during glacial cycles in the Southern Alpsof New Zealand, as revealed from OSL‐thermochronology, Earth Planet.Sci. Lett., 297, 183–189, doi:10.1016/j.epsl.2010.06.019.

Hetényi, G., R. Cattin, J. Vergne, and J. L. Nábělek (2006), The effectiveelastic thickness of the India Plate from receiver function imaging, grav-ity anomalies and thermomechanical modelling, Geophys. J. Int., 167(3),1106–1118, doi:10.1111/j.1365-246X.2006.03198.x.

Hetzel, R., and A. Hampel (2005), Slip rate variations on normal faults dur-ing glacial–interglacial changes in surface loads, Nature, 435, 81–84,doi:10.1038/nature03562.

Hodges, K. V. (2000), Tectonics of the Himalaya and southern Tibet fromtwo perspectives, Geol. Soc. Am. Bull., 112(3), 324–350.

Hodges, K. V., J. M. Hurtado, and K. X. Whipple (2001), Southward extru-sion of Tibetan crust and its effect on Himalayan tectonics, Tectonics, 20(6),799–809.

Hodges, K. V., C. Wobus, K. Ruhl, T. Schildgen, and K. Whipple (2004),Quaternary deformation, river steepening, and heavy precipitation atthe front of the Higher Himalayan ranges, Earth Planet. Sci. Lett., 220,379–389, doi:10.1016/S0012-821X(04)00063-9.

Hoth, S., A. Hoffmann‐Rothe, and N. Kukowski (2007), Frontal accretion:An internal clock for bivergent wedge deformation and surface uplift,J. Geophys. Res., 112, B06408, doi:10.1029/2006JB004357.

Huntington, K. W., and K. V. Hodges (2006), A comparative study ofdetrital mineral and bedrock age‐elevation methods for estimating ero-sion rates, J. Geophys. Res., 111, F03011, doi:10.1029/2005JF000454.

Huntington, K. W., A. E. Blythe, and K. V. Hodges (2006), Climate changeand Late Pliocene acceleration of erosion in the Himalaya, Earth Planet.Sci. Lett., 252, 107–118, doi:10.1016/j.epsl.2006.09.031.

Jouanne, F., J. L.Mugnier,M. R. Pandey, J. F. Gamond, P. Le Fort, L. Serrurier,C. Vigny, and J. P. Avouac (1999), Oblique convergence in the Himalayasof western Nepal deduced from preliminary results of GPS measurements,Geophys. Res. Lett., 26(13), 1933–1936, doi:10.1029/1999GL900416.

Kirby, S. H. (1983), Rheology of the lithosphere, Rev. Geophys., 21(6),1458–1487.

Kirby, S. H., and A. K. Kronenberg (1987), Rheology of the lithosphere:Selected topics, Rev. Geophys., 25(6), 1219–1244.

Konstantinovskaia, E., and J. Malavieille (2005), Erosion and exhumationin accretionary orogens: Experimental and geological approaches,Geochem.Geophys. Geosyst., 6, Q02006, doi:10.1029/2004GC000794.

Lachenbruch, A. H., and J. H. Sass (1980), Heat flow and energetics of theSan Andreas fault zone, J. Geophys. Res., 85, 6185–6222, doi:10.1029/JB085iB11p06185.

Larson, K. M., R. Bürgmann, R. Bilham, and J. T. Freymueller (1999),Kinematics of the India‐Eurasia collision zone from GPS measurements,J. Geophys. Res., 104, 1077–1094, doi:10.1029/1998JB900043.

Lavé, J., and J. Avouac (2000), Active folding of fluvial terraces across theSiwaliks Hills, Himalayas of central Nepal, J. Geophys. Res., 105, 5735–5770,doi:10.1029/1999JB900292.

Lavé, J., and J. P. Avouac (2001), Fluvial incision and tectonic uplift acrossthe Himalayas of central Nepal, J. Geophys. Res., 106, 26,561–26,591,doi:10.1029/2001JB000359.

Le Fort, P. (1975), Himalayas: the collided range. Present knowledge of thecontinental arc, Am. J. Sci., 275(1), 1–44.

Leloup, P. H., G. Mahéo, N. Arnaud, E. Kali, E. Boutonnet, D. Liu,L. Xiaohan, and L. Haibing (2010), The south Tibet detachment shear zonein the dinggye area: Time constraints on extrusion models of the Hima-layas, Earth Planet. Sci. Lett., 292, 1–16, doi:10.1016/j.epsl.2009.12.035.

Luttrell, K., D. Sandwell, B. Smith‐Konter, B. Bills, and Y. Bock (2007),Modulation of the earthquake cycle at the southern San Andreas fault bylake loading, J. Geophys. Res. , 112 , B08411, doi :10.1029/2006JB004752.

Lyon‐Caen, H., and P. Molnar (1985), Gravity anomalies, flexure of theIndian plate and the structure, support and evolution of the Himalayaand Ganga Basin, Tectonics, 4(6), 513–538.

Magee, M. E., and M. D. Zoback (1993), Evidence for a weak interplatethrust fault along the northern Japan subduction zone and implicationsfor the mechanics of thrust faulting and fluid expulsion, Geology, 21(9),809–813.

Maniatis, G., D. Kurfeß, A. Hampel, and O. Heidbach (2009), Slip accel-eration on normal faults due to erosion and sedimentation—Results froma new three‐dimensional numerical model coupling tectonics and land-scape evolution, Earth Planet. Sci. Lett., 284, 570–582, doi:10.1016/j.epsl.2009.05.024.

Martin, A. J., P. G. DeCelles, G. E. Gehrels, P. J. Patchett, and C. Isachsen(2005), Isotopic and structural constraints on the location of the Main


18 of 19

Central Thrust in the Annapurna Range, central Nepal Himalaya, Geol.Soc. Am. Bull., 117(7–8), 926–944.

Martin, A. J., J. Ganguly, and P. G. DeCelles (2010), Metamorphism ofGreater and Lesser Himalayan rocks exposed in the Modi Khola valley,central Nepal, Contrib. Mineral. Petrol., 159(2), 203–223, doi:10.1007/s00410-009-0424-3.

Nabelek, J., G. Hetenyi, J. Vergne, S. Sapkota, B. Kafle, M. Jiang, H. Su,J. Chen, B.‐S. Huang, and Hi‐CLIMB Team (2009), Underplating in theHimalaya‐Tibet collision zone revealed by the Hi‐CLIMB experiment,Science, 325(5946), 1371–1374, doi:10.1126/science.1167719.

Owen, L. A., and D. I. Benn (2005), Equilibrium‐line altitudes of the LastGlacial Maximum for the Himalaya and Tibet: An assessment and evalu-ation of results, Quat. Int., 138, 55–78, doi:10.1016/j.quaint.2005.02.006.

Owen, L. A., R. C. Finkel, and M. W. Caffee (2002), A note on the extentof glaciation throughout the Himalaya during the global Last GlacialMaximum, Quat. Sci. Rev., 21(1–3), 147–157, doi:10.1016/S0277-3791(01)00104-4.

Owen, L. A., M. W. Caffee, R. C. Finkel, and Y. B. Seong (2008), Quater-nary glaciation of the Himalayan‐Tibetan orogen, J. Quat. Sci., 23(6–7),513–531, doi:10.1002/jqs.1203.

Pandey, M. R., R. P. Tandukar, J.‐P. Avouac, J. Lave, and J.‐P. Massot(1995), Interseismic strain accumulation on the Himalayan crustal ramp(Nepal), Geophys. Res. Lett., 22(7), 751–754, doi:10.1029/94GL02971.

Parsons, T. (2002), Nearly frictionless faulting by unclamping in long‐terminteraction models, Geology, 30(12), 1063–1066.

Pearson, O. N., and P. G. DeCelles (2005), Structural geology and regionaltectonic significance of the Ramgarh thrust, Himalayan fold‐thrust belt ofNepal, Tectonics, 24, TC4008, doi:10.1029/2003TC001617.

Pratt, B., D. W. Burbank, A. Heimsath, and T. Ojha (2002), Impulsivealluviation during early Holocene strengthened monsoons, central NepalHimalaya, Geology, 30(10), 911–914.

Pratt‐Sitaula, B. (2005), Glaciers, climate, and topography in the NepaleseHimalaya, Ph.D. thesis, Univ. of Calif., Santa Barbara.

Rahaman, W., S. K. Singh, R. Sinha, and S. K. Tandon (2009), Climatecontrol on erosion distribution over the Himalaya during the past100 ka, Geology, 37(6), 559–562, doi:10.1130/G25425A.1.

Robert, X., P. van der Beek, J. Braun, C. Perry,M. Dubille, and J. L.Mugnier(2009), Assessing Quaternary reactivation of theMain Central Thrust zone(central Nepal Himalaya): New thermochronologic data and numericalmodeling, Geology, 37(8), 731–734, doi:10.1130/G25736A.1.

Robert, X., P. van der Beek, J. Braun, C. Perry, and J.‐L. Mugnier (2011),Control of detachment geometry on lateral variations in exhumation ratesin the Himalaya: Insights from low‐temperature thermochronology andnumerical modeling, J. Geophys. Res., 116, B05202, doi:10.1029/2010JB007893.

Robinson, D. M., P. G. DeCelles, C. N. Garzione, O. N. Pearson, T. M.Harrison, and E. J. Catlos (2003), Kinematic model for the Main Centralthrust in Nepal, Geology, 31(4), 359–362.

Schelling, D. (1992), The tectonostratigraphy and structure of the easternNepal Himalaya, Tectonics, 11(5), 925–943, doi:10.1029/92TC00213.

Schelling, D., and K. Arita (1991), Thrust tectonics, crustal shortening, andthe structure of the far‐eastern Nepal Himalaya, Tectonics, 10(5), 851–862,doi:10.1029/91TC01011.

Scherler, D., B. Bookhagen, and M. R. Strecker (2011), Spatially variableresponse of Himalayan glaciers to climate change affected by debriscover, Nat. Geosci., 4, 156–159, doi:10.1038/ngeo1068.

Schulte‐Pelkum, V., G. Monsalve, A. Sheehan, M. R. Pandey, S. Sapkota,R. Bilham, and F. Wu (2005), Imaging the Indian subcontinent beneaththe Himalaya, Nature, 435, 1222–1225, doi:10.1038/nature03678.

Searle, M. P., and L. Godin (2003), The South Tibetan detachment and theManaslu Leucogranite: A structural reinterpretation and restoration of theAnnapurna‐Manaslu Himalaya, Nepal, J. Geol., 111, 505–523,doi:10.1086/376763.

Searle, M. P., R. Law, L. Godin, K. Larson, M. Streule, J. Cottle, andM. Jessup (2008), Defining the Himalayan Main Central Thrust in Nepal,J. Geol. Soc., 165(2), 523–534, doi:10.1144/0016-76492007-081.

Stolar, D. B., S. D. Willett, and G. H. Roe (2006), Climatic and tectonicforcing of a critical orogen, in Tectonics, Climate, and Landscape Evolu-tion, edited by S.D.Willett et al., Spec. Pap. Geol. Soc. Am., 398, 241–250.

Thiede, R. C., B. Bookhagen, J. R. Arrowsmith, E. R. Sobel, and M. R.Strecker (2004), Climatic control on rapid exhumation along the SouthernHimalayan Front, Earth Planet. Sci. Lett., 222, 791–806, doi:10.1016/j.epsl.2004.03.015.

Tomkin, J. H., and G. H. Roe (2007), Climate and tectonic controls on gla-ciated critical‐taper orogens, Earth Planet. Sci. Lett., 262, 385–397,doi:10.1016/j.epsl.2007.07.040.

Tsenn, M. C., and N. L. Carter (1987), Upper limits of power law creep ofrocks, Tectonophysics, 136(1–2), 1–26, doi:10.1016/0040-1951(87)90332-5.

Turpeinen, H., A. Hampel, T. Karow, and G. Maniatis (2008), Effect of icesheet growth and melting on the slip evolution of thrust faults, EarthPlanet. Sci. Lett., 269, 230–241, doi:10.1016/j.epsl.2008.02.017.

Underwood, P. (1983), Dynamic relaxation, in Computational Methods forTransient Analysis, vol. 1, pp. 245–265, North Holland, Amsterdam.

Wessel, P., and W. H. F. Smith (1991), Free software helps map and dis-play data, Eos Trans. AGU, 72(441), 445–446.

Whipp, D. M., T. A. Ehlers, A. E. Blythe, K. W. Huntington, K. V.Hodges, and D. W. Burbank (2007), Plio‐Quaternary exhumation historyof the central Nepalese Himalaya: 2. Thermokinematic and thermochron-ometer age prediction model, Tectonics, 26, TC3003, doi:10.1029/2006TC001991.

Whipple, K. X. (2009), The influence of climate on the tectonic evolutionof mountain belts, Nat. Geosci., 2, 97–104, doi:10.1038/ngeo413.

Whipple, K. X., and B. J. Meade (2006), Orogen response to changes inclimatic and tectonic forcing, Earth Planet. Sci. Lett., 243, 218–228,doi:10.1016/j.epsl.2005.12.022.

Willett, S. (1999), Orogeny and orography: The effects of erosion on thestructure of mountain belts, J. Geophys. Res., 104(B12), 28,957–28,981, doi:10.1029/1999JB900248.

Wobus, C. W., K. V. Hodges, and K. X. Whipple (2003), Has focuseddenudation sustained active thrusting at the Himalayan topographicfront?, Geology, 31(10), 861–864, doi:10.1130/G19730.1.

Wobus, C., A. Heimsath, K. X. Whipple, and K. Hodges (2005), Activeout‐of‐sequence thrust faulting in the central Nepalese Himalaya, Nature,434, 1008–1011, doi:10.1038/nature03499.

Yin, A. (2006), Cenozoic tectonic evolution of the Himalayan orogen asconstrained by along‐strike variation of structural geometry, exhumationhistory, and foreland sedimentation, Earth Sci. Rev., 76(1–2), 1–131,doi:10.1016/j.earscirev.2005.05.004.

Zhang, P. Z., et al. (2004), Continuous deformation of the Tibetan Plateaufrom global positioning system data, Geology, 32(9), 809–812,doi:10.1130/G20554.1.

Zhao, W., K. D. Nelson, J. Che, J. Quo, D. Lu, C. Wu, and X. Liu (1993),Deep seismic reflection evidence for continental underthrusting beneathsouthern Tibet, Nature, 366, 557–559, doi:10.1038/366557a0.

Zoback, M. D., and G. C. Beroza (1993), Evidence for near‐frictionlessfaulting in the 1989 (M 6.9) Loma Prieta, California, earthquake andits aftershocks, Geology, 21(2), 181–185.

D. W. Burbank, Earth Research Institute, 6832 Ellison Hall, Universityof California, Santa Barbara, CA 93106‐3060, USA.V. Godard, CEREGE, CNRS, UMR 6635, Aix‐Marseille Université,

Europôle Méditerranéen de l’Arbois, BP 80, F‐13290 Aix‐en‐Provence,France. ([email protected])


19 of 19

mechanical analysis of controls on strain partitioning in the...

Documents